From: Michael Salfino
Date: Tue, Jan 15, 2013 at 12:22 AM
Subject: Breakfast of Champions
To: scott pianowski
That was one helluva interesting week of NFL football. And now we get to do it all over again on Sunday, but just twice.
Before getting to the championship games and predictions, we should close the book on Week 19. Who is the biggest goat in Denver: John Fox, Peyton Manning or the Broncos' secondary? Who would you rather have future shares on right now, Colin Kaepernick or Russell Wilson? Is Kaepernick going to burn out like Randall Cunningham did and even Kordell Stewart to a lesser extent, or is he going to endure? Here's a shocker - the Falcons were outplayed and won anyway. Why do I hate them for spitting in the face of football sabermetrics? And is there something about them that the numbers miss? Should Pete Carroll be given grief for not passing more early given how banged up Marshawn Lynch looked? And there are the Patriots, who played like champions on Sunday. But do we just give Bill Belichick a pass on losing Rob Gronkowski? Does Gronk's loss even matter given that no one has a clue how to stop the Patriots from doing the same stuff they've been doing, seemingly, forever?
And why does the indoor game happen early and the outdoor game get played late? Plus they're making the west coast team defy circadian rhythms again a week after you could almost set your clock to the minute the Seahawks would wake up and play. Ratings, I know, but they're both going to be so high anyway. I guess more people will want to see the Ravens come into Foxboro confident they can win given they beat New England earlier this year and should have beaten them last year in the same AFC Championship spot.
But I am looking forward to both games and am chastened, appropriately, after last week's predictions gone awry. Screw the bookies and pass those martinis. Breakfast of Champions is served.
From: scott pianowski
Date: Wed, Jan 16, 2013 at 9:42 PM
Subject: outfoxed *use this version*
To: Michael Salfino
I'm glad John Fox has never been tied to a team I care about. I hope that's always the case. There's your goat. Got no spine, coaches orange crush.
On two separate occasions at the end of a half, the Broncos decided to scrap the freeroll and simply let the clock run out. Unless there was some sort of physical problem with Peyton Manning that hasn't been disclosed, this is inexcusable. The chances of Denver scoring in those instances far outweigh the likelihood of a mishap and a Baltimore score. But if there's ever been a coach who has "Play Not To Lose" tattooed to his forearm, it's John Fox. (You'll note that on the same weekend, Houston and Atlanta put together quick drives with limited time. The Falcons did have desperation fueling the march, I suppose.)
Here's a simple way to judge most game-theory decisions: ask what a reasonable opponent would want you to do in that instance. Turn it into a game of prisoner's dilemma. I'm sure the Ravens were thrilled to see Manning's kneel-down as the fourth quarter expired.
If John Fox were in Delta House, the movie ends when everyone gets expelled. John Fox didn't take a big chance at any high school dance. I bet John Fox never took a lead off any base during his hardball career. Good luck with this guy, Denver.
Mind you, it wasn't the best weekend for the head coaches. Pete Carroll and Mike Smith had a Dumb and Dumber marathon in their game. Look, I respect and appreciate how difficult coaching football is at any level. The Xs and Os component of the NFL is ridiculously complicated, and would go over the head of 99 percent of the fans in the stands. Their job is harder than it looks, and decision making is harder in the heat of the moment, bullets flying. We're all Jeopardy! champions in our own home, on the couch with nothing at stake.
But I'm blown away at how many coaches don't understand the basics of game theory and, specifically, modern football game theory. And I don't think it's a big coincidence that three of the coaches inside the Age of Enlightenment, Belichick and the Harbaugh brothers, are still in the tournament.
Rob Gronkowski's injury definitely matters. He's far and away the best tight end in the league, and New England's second-most important player. Danny Woodhead is a loss, too. If New England had all of its offensive pieces in place for the full season, this could have been the best unit of all time. Heck, they were in striking distance of the scoring record anyway, and that's with Gronk and Aaron Hernandez missing several games between them.
A huge part of Gronk's value comes through in the trickle-down effect - what he forces a defense to do, and how it enables others to have cushy matchups. Contrast this to the New England defense, where the acquisition of Talib has allowed McCourty to blossom as a safety.
I'll let you take a more detailed look back, and then have the first look ahead. I'm probably picking New England (though the line is ridiculous) and San Francisco, but I want one last look at the tapes.
From: Michael Salfino
Date: Thu, Jan 17, 2013 at 11:15 AM
Subject: Room Full of Mirrors
To: scott pianowski
How can the coach be the goat when the team had the lead with 40 seconds left and lost on a 70-yard bomb that was the worst coverage imaginable at the worst possible time? The Broncos win probability before that play was 99%. And the key play of the game no one talks about was the 24-yard pass to Dennis Pitta on third and 13 from their own three in overtime; prior to that play, the Ravens had a 29 percent chance of winning. The Pitta catch, thanks Jim Leonhard, swung the odds of a Ravens victory about more than half as much as the Jacoby Jones TD.
If I'm Baltimore, I want the Broncos to try to move the ball 50 yards in 30 seconds against me. The Broncos are not a downfield passing team and thus are poorly designed to get big chunks of yards through the air. That's their B game, at best. Probably C or D. Only two of Manning's 43 passes went over 15 yards in the air (NFL average would be 8-9). So I don't have a problem with running the clock down before the half or at the end of the game, given their field position both times. I also believe they were against the wind both times. Plus at the end of the game, you need a standing eight count there, anyway. Bottom line for me is that they had a 99 percent chance to win in regulation and a 71 percent chance to win in overtime, so it's tough to unwind the game from those two points and say the Fox is an idiot. Running the ball before the Ravens TD was a 100 percent right call, and anyone who would risk a pass there is the idiot.
I know from our talks backstage that you have a problem with win probability. I have no idea why. You often cite the value of field position all the time, and that's based on exactly the same thing. This idea that we're going to play the game based on what we think the opponent thinks he doesn't want us to do is good shorthand, but pretty soon you are in a room full of mirrors (cue Jimi). I'd rather just know how key decisions impact my chances of winning based on what's happened in the same down and distance, score and time remaining in the relevant history. Then I can adjust that to what I think my team does well and the opponent poorly, relative to average. Do you think the Broncos are a better-than-average downfield passing team?
And if you love the Harbaughs so much, to be consistent, you should be bashing John for not going for two and the win after the Jones TD. Why give Peyton Manning a chance to beat you at home in the high altitude when you have a better than 50 percent chance to convert and win right there? I sense outcome bias at work here. Oh, and Jim Harbaugh is the genius who can't figure out a way to get Vernon Davis the ball while still feeding the ghost of Randy Moss.
What did Carroll and Smith do incorrectly in their games? Going for it on fourth and 1 at the Falcons 11 was the right call by Carroll though the play call itself sucked. However, I don't know if that was a decision made by Wilson. And the Seahawks rushed Ryan at the end, which I think was stupid but runs contrary to the things you are criticizing Fox about so I assume you liked it. But what did Smith do wrong? Nothing stands out for me, plus Smith, according to Massey-Peabody at least last year was the best coach in the league at understanding win probability based on going for it on fourth down, eschewing field goals, etc.
I don't like the 49ers playing another early game. What is so complicated about this, NFL? Don't screw the West Coast teams in games like this by having them kick off before 1 p.m. PST. But the idea that the Falcons are going to try to attack Kaepernick is foolish. There's also talk of heavy blitzing that seems likely to be sourced to a locker-room contact. This is a huge mistake. The way to play Kaepernick ironically is the same way you play a big-time passing team. Be passive. Mush rush. Don't even try to sack him. Drop seven or eight guys back into a zone with their eye on him on hold your ground until he advances past the linemen, then you can attack. But teams hate doing this because when you are a hammer, every problem is a nail. Instead of blitzing, you want coverage sacks where he's spinning around and reversing field in the wrong direction. You need total control and discipline on every snap, counting on your teammates to make a play and not trying to be a hero and taking yourself out of position. Teams generally can't do this, though, which makes Kaepernick so good, for now. When they see someone do it though, they will all copy it. That's what ultimately got the better of Stewart and Cunningham. The Falcons offense is going to score 20-plus points at home, though. I was impressed by them versus a Seahawks defense that I think is better than San Francisco's. This game can end up being a classic. Expect Atlanta to get off to a fast start again. But I think the Niners close it out, 27-24.
The Patriots obviously have trouble with the Ravens. Now Baltimore, with new offensive coordinator Jim Caldwell (who so many people just hated), suddenly has a good offense but not nearly a typical Ravens defense. I do not have confidence in New England to stop anyone. Flacco is a tough matchup because he throws the ball so high and far, and teams really don't do that anymore so these defensive backs seem to have trouble judging the ball. One coach told Phil Simms he had the scout team throw them as high as possible and when you add up catches and pass interferences, the starters allowed something like 14-for-20 at practice. And as you know, there isn't a lot of practice time anymore. I can't believe you are talking about Patriots mascot Danny Woodhead like he's some key player. Shane Vereen is clearly better in every way. Woodhead, by the way, had ProFootballFocus's worst broken tackle/elusiveness rating out of 52 backs, one spot worse than Shonn Greene.
He is on the team only so that writers can glorify the ingenuity of Bill Belichick in finding players like Woodhead off the scrapheap, never mind that every snap for him keeps better players off the field. But the Patriots are generally coached so freakin' well. They run those two plays on the goal line that no one can stop - put your tackles in the A gaps, and the Patriots sweep the back untouched into the endzone. Don't cover those gaps and Brady just runs a sneak. And it's all a fire drill where the defense can't even get set up. And Brady is still so clearly at the top of his game. Plus he gets about six or seven free passes every game where a guy isn't covered or the defense is scrambling to get into position. I just can't bet against him, even though the New England defense is terrible. Patriots 35, Ravens 31.
From: scott pianowski
Date: Thu, Jan 17, 2013 at 9:46 PM
Subject: too long, didn't read
To: Michael Salfino
I've been torpedoed by the Boogie Woogie Flu all day, prey to whatever's going through the Midwest. Peyton Hillis told me to bag it, but we play through.
Now, back to the argument.
I get what you're saying about Denver's secondary. Obviously Tony Carter and Rahim Moore made horrible errors in judgment on the Jacoby Jones touchdown; that play has the feel of a Denver giveaway more than a Baltimore accomplishment. And Champ Bailey was schooled all afternoon, a surprising (maybe even shocking) turnaround from his fine play all year. These guys deserve all the flags and blame they're getting.
But I guess in this post mortem I want to focus on strategic slants, not simply pointing at a unit and saying "hey, you guys, play better - and stop misjudging the ball while it's in the air." Execution comes and goes, on-field judgment errors come and go. Flaws in game theory stick around.
Here's what I challenge anyone to find: I'd like to see proof that if the Broncos keep the ball in play in the final half-minute of regulation *it favors the Ravens* from a point-expectancy perspective. Maybe it's convenient for me to frame it that way; I suppose someone could challenge me to prove it's a better spot for the Broncos (or for any offense). With two timeouts in pocket and an offense led by Peyton Manning, it seems obvious to me (though I have no data-driven proof) that the Broncos will win that game in regulation far more often than the Ravens will. But maybe I'm missing some study somewhere, maybe I'm obtuse to some key factor.
(As for Denver not being a quick-strike offense, we're not asking them to escape from Shawshank Prison in 30 seconds here. If they simply click on a few medium plays, a kick is doable - especially in thin air. Heck, just call an upside play with low risk, say a screen, and see what it does for you. If it doesn't work, fine, call off the dogs.)
I wish you hadn't brought up the Win Expectancy comment out of context, an off-hand statement I made during the unofficial backstage table with our buddy Steve Moyer, but I guess this will turn into a public chat. Here's exactly what I said (and I understand this will be presented without the context of the extended discussion between the three of us, but that's how the cookie crumbles):
I look forward to the future date when we look back and laugh at how inaccurate Win Probability was in the early years.
Oh, wait, I forgot, today's statistical inroads are an answer key.
WAR runs the world!
Maybe I assume longtime friends like you and Steve will know when I'm stretching a point or being partially sarcastic, even in an email - there's plenty of intentional exaggeration in this passage - but there's also a base point at the heart of this note that I do believe in. Perhaps I should just go back to Gene McCaffrey's elegant roto observation from his 2011 annual (seriously, everyone, read Gene McCaffrey whenever you can):
Ten years ago nobody ever heard of BABIP, now it's as if nothing else matters. But this is good for us, because it locks otherwise intelligent people into not thinking things through.
Gene's the best. And that's how I feel about any "one number" view of the world. Heck, there's a tangible gap between Baseball-Reference WAR and Fangraphs WAR (which have both moved away from the original Bill James concept), and defensive evaluation remains a big, fat mess. Remember how important OPS seemed to be a decade ago? We've moved on to better, greater things. You did some wonderful work last year with respect to how BABIP tells an incomplete story, especially for a pitcher; you started asking the right questions on Johnny Cueto before (as far as I know) anyone else.
My main takeaway is that intelligent theory will always evolve, and whatever's around in 10, 20, 40, 60 years will make today's work in large doses obsolete. And I guess I wanted to take a swipe at some of the baseball scribes who consider WAR to be the gavel towards any argument; move along, have a nice day.
Of course the football statheads have a much more difficult task on their hands, because football is a context-driven game - contrast that to the largely-individual nature of baseball. No one needs to help Miguel Cabrera clobber a home run (and he's going to clobber them in any context), but a touchdown pass from Flacco to Smith involves 20 other players, not to mention the coaches and maybe even the officials. Football's complicated, man. You take one Lego out of place and everything potentially collapses.
I would have been fine with Jim Harbaugh going for the 2-pointer after the Jones touchdown. I was discussing this with my poker buddies right after the score. The nifty thing about that move is that you have a backboard if you fail on the deuce - you can still try an onside kick. If the visiting team felt it were at a significant talent disadvantage - I don't think it applies in this game - I'd argue the conversion is definitely the right play. And I hope someday pro coaches, especially those in a no-pressure situations, are more inclined to try this sort of gambit. When two 4-11 teams play in Week 17, why not empty out the gimmick shelf and see what works and what doesn't?
The first error Mike Smith made came after the Falcons scored late in the third period to take a 26-7 lead (kick pending). Atlanta probably should have gone for two no matter what (being up 21 is better than 20 at that stage of the game), but the Seahawks augmented the case by jumping offside on two consecutive kicks (fill in your own milligram joke here), pushing the ball closer to the goal. The gain of a 20-point lead over a 19-point lead is that you no longer have to sweat the 8+8+3 tie game - two touchdowns with conversions and a field goal. Page me when that happens someday. The upside of being up 21 is that you are now tied against three sevens. Missed opportunity, Atlanta. And it almost cost you the ballgame.
I suppose that's a minor error; a lot of coaches are oblivious to the 2-point theory in games that don't appear to be close. But it's still an oversight.
Atlanta also mishandled the clock on the game-winning field goal drive, though it's easy to see how it happened. The Falcons logically burned their final timeout after the Gonzalez catch put the ball down at the Seahawks 31-yard-line. It takes too much time to run a spike then, so the automatic timeout is the play - unless you already know that you're going to settle for the resulting 49-yard kick. Had the Falcons thought that quickly on their feet (and I can't blame them for not being able to crunch it immediately), letting a few seconds run off the clock before the timeout is the right move, give Seattle no chance at a tangible comeback.
What the Falcons could have done, and probably should have done, is run one (maybe even two) more plays, with Ryan dropping back and immediately firing a sideline pass. If it's not there, fine, chuck it into the next county. Drain some clock. Maybe you get lucky and move up a few yards. Banking on a 49-yard make for your playoff life is not how I want to live, but if you're going that route, don't leave the back door open for the other guy.
Seattle's big regret is how conservatively it ran the offense in the first half. Russell Wilson needed to be unleashed earlier. But we also have to look at Pete Carroll's silly ice-the-kicker timeout, a move that has no logical foundation for anyone attempting a long kick. If you want to play with the opponent's head on a short try, I can see it. Hell, you might mess with a kicker's mojo more if you don't ice him, since he might be thinking in the back of his mind that he's going to get iced. But all Carroll's gambit did was effort Bryant a practice kick.
Atlanta then graciously allowed Seattle a viable comeback with a horrendous squib kick (I suppose that's more of an execution problem than a strategic one; Leon Washington is a threat) and an easy sideline pass. If the Seahawks had a kicker with a long leg, a 66-yarder is possible indoors. Wilson's Hail Mary made it to the end zone; alas, more Falcons were ready for the pass and Julio Jones secured it rather easily. Where are Duriel Harris and Tony Nathan when you need them?
All this rhetoric and I've left the fresh games alone. I look forward to the tomato throws from the peanut gallery.
Perhaps Atlanta's biggest problem is the fact that it hasn't faced San Francisco yet. It's hard to prepare for San Fran's speed and physicality merely on tape; there's an adjustment period of sorts. Consider how the New England-San Francisco game played out. Also consider that every NFC West team did better in its rematch with the Niners (which included a Seattle romp and a stunning Rams win). The location of the game favors the Falcons, but the surface helps the Niners. San Francisco 27, Atlanta 23.
I don't think the Patriots defense is as bad as you think, but Flacco's done very well in this series (last three: 8.7 YPA, 114.8 rating, 8 TDs, one pick). While it's absurd for some to suggest New England would have been better off playing at Denver this week, Baltimore is the second-hardest potential conference matchup for New England. Obviously the Ravens have won here before in the playoffs (and should have won a year ago), and they also beat the Pats in that messy Week 3 officiating mess. The line's a joke, too - New England's a public team, kids.
I'm calling it Patriots 31, Ravens 27, and I give Baltimore about a 35-40 percent chance to win. New England needs to avoid another key loss on offense, and obviously it can't turn the ball over (you can say that about any team, any week). But I figure the Patriots will eventually solve this ordinary Baltimore defense, and once that happens, the points flow easily. Ask J.J. Watt, he'll tell you. The Pats don't need a ton of stops on defense, just one more than the Ravens get.
Enjoy the games, amigos.